Abstract:In order to improve sensitivity and noise immunity of a single index to damage in damage identification, based on modal strain energy theory, a two-level feature fusion method combining principal component analysis and entropy weight method are proposed. The Northern Goshawk Optimization (NGO) algorithm combined with Gated Recurrent Unit (GRU) are used for bridge damage degree prediction. Firstly, based on traditional modal strain energy theory, the diagonal modal strain energy ratio is constructed, and then change rate of the diagonal modal strain energy ratio, dissipation rate of the diagonal modal strain energy ratio, and normalized difference index of the diagonal modal strain energy ratio are derived. Secondly, principal component analysis is used to extract features within the index, and entropy weight method is used to fuse features between indexes. Finally, Weighted Decision Index (WDI) is constructed. The single modal strain energy derivative index is input into the NGO-GRU hybrid neural network, as well as damage degree is output, so as to establish the relationship between index value and damage degree, and then realize damage quantification. The method proposed in this study was verified by a three-span continuous beam bridge numerical model. The results show that weighted decision index has good damage location ability and noise immunity. The hybrid neural network has high damage prediction accuracy, with a prediction accuracy rate of 91.14%.